Please enable JavaScript.
Coggle requires JavaScript to display documents.
Artificial Intelligence, Readability and quality of AI-generated feedback,…
Artificial Intelligence
- Artificial Narrow Intelligence (ANI) or Weak AI -
- Artificial General Intelligence (AGI) -theoretical
- Artificial Superintelligence (ASI)
Education Context
Trends
-
2018 Horizon Report - identified AI and adaptive learning as emerging technologies, forecasting 43% growth between 2018 and 2022.
STUDIES
GPT-3 reliably scored 12,100 TOEFL11 essays, with improved accuracy when combined with linguistic features. Their findings suggest that AI can effectively support human raters and enhance writing assessment practices.
ChatGPT as a tool for providing student feedback. Recognizing the role of effective feedback in learning, the study assessed ChatGPT’s ability to produce natural, comprehensive responses. ChatGPT offered more detailed and coherent feedback, aligned well with instructor evaluations, and supported students’ skill development by focusing on task performance.
ChatGPT excels in addressing specific language-related issues such as grammar, spelling, and sentence structure, showcasing its proficiency in rule-based corrections. However, human evaluators maintain an edge in areas requiring nuanced understanding, particularly in maintaining thematic consistency
Results showed no significant difference in learning outcomes, and students expressed balanced preferences for both. The findings suggest that AI feedback can complement writing instruction when paired with human input, enhancing the strengths of each.
Among them, closed-source models like GPT-4 and o1 showed stronger alignment with teacher evaluations, especially on language aspects. While o1 had the highest correlation, it tended to overrate. The study suggests that LLMs can aid in writing assessment but still require refinement for evaluating content quality.
EvaluMate, an AI-assisted peer review platform using ChatGPT to help students give more effective feedback. The tool offers scaffolding to improve comment quality, showing promise for writing instruction. While beneficial, the study also notes limitations and suggests further research on student engagement and learning impact.
Histororical Timeline
Initial Launching - on November 30, 2022, ChatGPT attracted over one million users within just a week. This generative AI tool surprised the world with its advanced ability to handle highly complex tasks. In the realm of education, ChatGPT's impressive capabilities have sparked a range of reactions among educators, as this technological breakthrough appears to transform traditional educational practices.
Recent years have seen a major development in natural language processing (NLP) thanks to large language models like the Generative Pre-trained Transformer (GPT-3) . These models can produce text that is human-like, respond to inquiries, and perform other language-related tasks with great accuracy since they have been trained on vast volumes of text.
In the context of education, AI has become a debatable topic regarding its integration in teaching and learning. While academic institutions worldwide engage in discussions about AI’s role in assessment, the focus often centers on concerns about academic integrity, rather than its ability to improve the reliability of the assessment process.
AI research began with John McCarthy, who coined the term at a 1965 Dartmouth conference.
Significance
It has been frequently asserted in recent years that artificial intelligence (AI) is the "new oil" or, as the UNESCO Director General stated in her keynote address for Mobile Learning Week in 2019, the greatest invention since the Paleolithic era Link Title
-
Definitions
AI as machine-driven systems that make predictions, suggestions, or decisions based on human-defined goals.
-
-
-
-
-
-
-
Ethical and effective use of generative AI tools,in supporting student writing
-
-
-